eginning with the 2003 Tucson gem shows,cesium-rich “beryl” from Ambatovita,Madagascar, created excitement among gem
collectors and connoisseurs due to its deep purplishpink color (figure 1) and the attractive chatoyancydisplayed by some of the material. Although it wassold during and after the Tucson gem shows as“raspberyl,” “red beryl,” “pink beryl,” “hot pink-red beryl,” and especially “raspberry beryl,” subse-quent examinations (see, e.g., Simmons et al., 2003)revealed properties that were anomalous for beryland were associated with very high concentrationsof cesium (Cs). In September 2003, the Inter-national Mineralogical Association approved thename pezzottaite (pe-zó-ta-ite) for this new speciesof the beryl group. It is named after Dr. FedericoPezzotta (Natural History Museum, Milan, Italy),who was among the first to investigate this newmineral, in recognition of his scientific contribu-tions to the mineralogy of Madagascar (Hawthorneet al., 2003).
With the addition of pezzottaite—ideallyCs(Be2Li)Al2Si6O18—the beryl group consists of fourmembers. The other three (all hexagonal) are: beryl(Be3Al2Si6O18; Aurisicchio et al., 1988), bazzite
(Be3Sc2Si6O18; Armbruster et al., 1995), and stoppaniite(Be3Fe2Si6O18; Ferraris et al., 1998; Della Ventura etal., 2000). Pezzottaite, which is rhombohedral, isnot a Cs-rich beryl but rather a new mineral speciesthat is closely related to beryl. Another mineral,indialite ([Al2Si]Mg2[Al2Si4]O18; Meagher and Gibbs,1977), is also sometimes included in the berylgroup. Of these, only beryl and pezzottaite havebeen found in gem quality. (Note: In this article,“beryl” refers to the mineral species, rather than thegroup, unless otherwise specified.)
Pezzottaite has been confirmed from just onedeposit in Madagascar, the Sakavalana granitic peg-matite located near Ambatovita in a remote area ofthe central highlands. In addition to material fromthis locality, a sample of Cs-rich “morganite” fromAfghanistan described by Hänni and Krzemnicki(2003) has now been recognized as pezzottaite (H.Hänni, pers. comm., 2003). Observations of thecrystal morphology of pezzottaite from Madagascar
284 PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003
B
PEZZOTTAITE FROM AMBATOVITA,MADAGASCAR: A NEW GEM MINERAL
Brendan M. Laurs, William B. (Skip) Simmons, George R. Rossman, Elizabeth P. Quinn, Shane F. McClure, Adi Peretti,Thomas Armbruster, Frank C. Hawthorne, Alexander U. Falster, Detlef Günther, Mark A. Cooper, and Bernard Grobéty
Pezzottaite, ideally Cs(Be2Li)Al2Si6O18, is a new gem mineral that is the Cs,Li–rich member of theberyl group. It was discovered in November 2002 in a granitic pegmatite near Ambatovita in cen-tral Madagascar. Only a few dozen kilograms of gem rough were mined, and the deposit appearsnearly exhausted. The limited number of transparent faceted stones and cat’s-eye cabochons thathave been cut usually show a deep purplish pink color. Pezzottaite is distinguished from beryl byits higher refractive indices (typically no=1.615–1.619 and ne=1.607–1.610) and specific gravityvalues (typically 3.09–3.11). In addition, the new mineral’s infrared and Raman spectra, as wellas its X-ray diffraction pattern, are distinctive, while the visible spectrum recorded with the spec-trophotometer is similar to that of morganite. The color is probably caused by radiation-inducedcolor centers involving Mn3+.
were presented by Warin and Jacques (2003), andthe chemical composition was investigated byAbduriyim and Kitawaki (2003). This article pro-vides information on the history, geology, composi-tion, and properties of pezzottaite, as a follow-up tothe initial description of the material provided bySimmons et al. (2003). A separate article that willformally describe the new mineral (Hawthorne etal., in preparation) and will be submitted to theMineralogical Record.
HISTORYAs recounted by Dr. Pezzotta (pers. comm., 2003),mining of the Sakavalana pegmatite (for tourmaline)by French colonists began in the 1940s. Sub-sequently, intermittent digging by local people withhand tools occasionally produced tourmaline andother minerals. In mid-November 2002, the localminers discovered a large crystal-bearing cavity thatcontained multicolored tourmaline prisms of carv-ing or slabbing quality, as well as gem-quality spo-dumene crystals in green, blue, and purple hues.The production eventually reached the capital cityof Antananarivo, where it was seen by gem dealerLaurent Thomas in early December 2002. Henoticed some unusual deep pink crystals with acolor that resembled tourmaline but with a mor-phology similar to tabular morganite. Mr. Thomasmeasured a refractive index of 1.619 on a shinypinacoidal face on one of these crystals. Recognizingthat this was too high for beryl, he sent samples toDr. Pezzotta in Italy.
Preliminary chemical analyses of two sampleswith energy-dispersive X-ray spectrometry (by Dr.
Alessandro Guastoni of the Natural HistoryMuseum, Milan) revealed very high concentrationsof Cs, and calculation of the unit-cell parameters (byDr. Franco Demartin of the University of Milan)yielded data consistent with a beryl-like mineralwith extreme Cs enrichment. After the 2003Tucson gem shows, the customs department ofMadagascar temporarily froze the export of thematerial due to confusion over its identity and the rumor that it might be a new mineral.Subsequent studies proved that it was, indeed, anew mineral (see, e.g., Laurs, 2003; “Newly discov-ered beryl . . . ,” 2003).
Some of the initial production was mistakenlysold in Madagascar as tourmaline (F. Pezzotta, pers.comm., 2003). Soon, rumors that it represented anew source of red beryl (such as that from Utah)spurred local traders to call it “bixbite,” after theoutdated trade term for such material. Due to theexcitement of the discovery and strong demand,dealers frequently have encountered exorbitantprices for gem rough and crystal specimens, even inMadagascar.
LOCATION AND ACCESSDr. Pezzotta guided the senior author to the minein July 2003. It is located about 140 air-km (87miles) southwest of Antsirabe, in centralMadagascar (figure 2). Paved roads lead fromAntsirabe to Ambatofinandrahana, a distance ofapproximately 160 km that requires about sixhours of driving time. From there, a rough dirttrack proceeds approximately 140 km to the townof Mandrosonoro, and then another 25 km to the
PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003 285
Figure 1. Pezzottaite, anew gem mineral from
Madagascar, is aCs,Li–rich member ofthe beryl group. Small
mine. The assistance of local guides was necessaryfor both navigational and safety reasons. This driveusually takes about 14 hours, although we encoun-tered considerable delays due to multiple vehicularbreakdowns caused by the rough terrain (figure 3).Our excursion was undertaken in “ideal” condi-tions during the dry season; the road is hazardousto impassible during the rainy season, fromDecember to April.
The mine is situated on a low hill (figure 4), atcoordinates 20° 44.78’ S and 46° 04.45’ E and anelevation of 920 m (3,020 feet). It is located a fewkilometers northwest of the village of Ambatovitaand southwest of the village of Ankosira, both ofwhich lie along the Manambaroa River. Most ofthe miners live on-site, in several huts above theworkings (figure 5).
GEOLOGYThe Sakavalana pegmatite, which hosts the pez-zottaite, is located in the northern part of thefamous Ampandramaika-Malakialina pegmatitedistrict, one of many areas in central Madagascarthat have produced gem minerals (see Pezzotta,2001). A brief review of the regional geology ofthese pegmatites was provided by Dirlam et al.(2002). The geology of the Mandrosonoro-Ambatovita area was described and mapped byChantraine (1966). The Sakavalana pegmatite ishosted by impure marbles that Chantraine (1966)assigned to the Vohimena Group, on the easternlimb of a synform that is composed mainly ofquartzitic gneisses (again, see figure 2). Thisamphibolite-grade metasedimentary sequence issurrounded mostly by deformed granites.Pegmatites occur sparingly in the synform, but aremuch more abundant within mica schists thatform the upper facies of the Vohimena Group,about 20 km to the west. This area (Malakialina)
286 PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003
Figure 3. Mechanical breakdowns were commonbecause of the rough terrain, which required the use ofa high-clearance four-wheel-drive vehicle. Photo byBrendan Laurs.
Figure 2. The pezzottaite deposit is located in centralMadagascar, about 140 air-km southwest of
Antsirabe and 130 km northwest of Fianarantsoa.From Antsirabe, a paved road leads south to
Ambositra and west to Ambatofinandrahana, fol-lowed by a rough dirt track that passes through
Amborompotsy and Mandrosonoro. The pegmatite ishosted by marbles, on the limb of a syncline that is
composed mainly of quartzitic gneisses of theVohimena Group. Geology from Chantraine (1966).
PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003 287
once was the most important producer of industri-al beryl in Madagascar (Pezzotta, 2001).
Pezzottaite occurs at Sakavalana within a “mixedfeature” granitic pegmatite—that is, one that hascharacteristics of both the LCT (lithium, cesium,and tantalum) and NYF (niobium, yttrium, and fluo-rine) families of the Rare-Element and Miaroliticclasses (see `Cerny, 1991). The pegmatite (figure 6) isa subvertical dike that measures at least 4–6 m wideand more than 200 m long. The outer portion con-sists mostly of K-feldspar, quartz, plagioclase, andblack mica, whereas the core zone consists of K-feldspar (green amazonite; figure 7), zoned crystals ofblack and purple mica, and smoky quartz, withtraces of albite (bladed “cleavelandite” aggregates),danburite, zircon, and Nb-Ta oxides. In places, blacktourmaline is intergrown with these minerals in thecore zone, along with minor beryl, spessartine, andspodumene. Crystal-lined cavities are also locallypresent in the core zone, and they contain the core-zone minerals as well as pezzottaite in places. Theparagenetic relations of the pocket minerals indicatethat pezzottaite crystallized from the fluids in thecavities as late-stage crystals. Observations ofnumerous mineral samples by Dr. Pezzotta andsome of the authors indicate that post-crystallizationfluids caused etching or extensive corrosion onmany of the pezzottaite crystals.
The main pezzottaite find occurred about 6 mbelow the surface, within a large cavity that mayhave measured up to 3.0 × 3.0 × 1.2 m. Below thispocket was a zone containing numerous vugs withsmaller crystals of pezzottaite. All the pezzottaiteproduced to date came from this rather limitedarea within the pegmatite.
MINING AND PRODUCTIONThe mine area consists of pits, shafts, and opencuts that explore at least one steeply south-dippingpegmatite dike. These workings were probably dugonly with hand tools, although bulldozers avail-able in the early days of mining may have beenused. Illumination in the underground workings isby candlelight, and logs are used as crude scaffold-ing near the opening of the deepest shafts (approxi-mately 30 m). Flooding is a problem in the lower-most shaft (which produced only tourmaline). Theunderground workings where the pezzottaite wasmined (see, e.g., figure 8) consist of two smallrooms, which are accessed from above and belowby tunnels that are several meters long.
In the rush that followed the original find,approximately 120 people (including 60 miners andtheir families) lived at the deposit. Most of themleft the area in April 2003 due to lack of produc-tion. As of July 2003, only about 20 people wereactively working the pegmatite. Most of the pez-zottaite-bearing zone appeared to have been mined
Figure 5. At the pezzottaite mine, most of the minerslive on-site in several huts above the workings. One ofthe entrances to the pezzottaite-bearing part of the peg-matite is visible on the lower right (without a signifi-cant mine dump below it). Photo by Brendan Laurs.
Figure 4. In this view (looking southwest) of the pez-zottaite mining area, on the left are some of the min-ers’ huts and on the right is a small open cut that hasbeen worked for multicolored tourmaline since the1940s. Photo by Brendan Laurs.
out, and the remaining areas were difficult to workwith the hand tools available. Small amounts ofpezzottaite were available at the mine, as well aselsewhere in Madagascar, but the material shown
to us was of low quality. After another visit to themine in November 2003, Dr. Pezzotta reported thatthe underground workings were in poor conditiondue to rockfalls and backfill generated by contin-
288 PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003
Figure 7. The core zone of the pegmatite is marked byconcentrations of green amazonite. Note also the
large crystals of black tourmaline above the miner.Photo by Brendan Laurs.
Figure 9. This pezzottaite specimen has been cleanedof dirt and clay, but otherwise shows the appearanceas recovered from the mine. The pezzottaite crystalmeasures 3.6 cm in diameter, and is associated with“cleavelandite” feldspar (bladed albite, here partiallycovered with Mn-oxides) and quartz. Courtesy ofFederico Pezzotta; photo by Brendan Laurs.
Figure 6. The Sakavalana pegmatite is visible in thisopen cut as the light-colored dike on the left. The
miners are standing in front of a 13-m-deep shaft thatproduced multicolored tourmaline crystals. The
entrance to a separate tunnel that leads upward tothe pezzottaite workings is visible behind the man on
the far left. The green color of portions of the peg-matite and mine dump is due to the presence of ama-
zonite feldspar. Photo by Brendan Laurs.
Figure 8. Numerous pezzottaite-bearing vugs weremined from this area, using simple hand tools andcandlelight. Fractures in the pegmatite are coatedby iron-stained clay, which obscures observation ofthe mineralogy and also causes stability problemsin the ceiling, resulting in unsafe conditions. Photoby Brendan Laurs.
ued mining activities. Apparently this work pro-duced little (if any) additional pezzottaite.
According to Dr. Pezzotta (pers. comm., 2003),the main pezzottaite pocket produced a total ofabout 700 kg of smoky quartz (in crystals up to 50kg), 280 kg of tourmaline (multicolored, with ablack “skin”), >40 kg of pezzottaite (see, e.g., figure9), and 25 kg of transparent spodumene (see, e.g., fig-ure 10). The three largest tourmaline crystalsweighed 15, 25, and 50 kg, and reportedly were pur-chased by a Korean dealer for carving purposes. Thepezzottaite was recovered as crystals, fragments,and masses typically weighing a few grams, andrarely up to 1,000 grams. In addition to the 40 kgmentioned above, there were many tens of kilo-grams of low-quality fragments and deeply corrodedcrystals. Pieces of gem-quality pezzottaite are rareand typically small (less than 1–2 grams). Much ofthe gem material is of carving quality only,although some contains abundant tubes orientedparallel to the c-axis that produce attractive chatoy-ancy in cabochons. One gem dealer estimated thatperhaps 10% of the rough will produce the cat’s-eyeeffect when cut (“Rare pink-red beryl. . . ,” 2003).
The total amount of pezzottaite produced so farfrom the pegmatite is difficult to estimate. Accordingto dealers Fabrice Danet, Denis Gravier, DudleyBlauwet, and Mark Kaufman (pers. comms., 2003),the total production may range up to 150 kg, with nomore than 25% containing areas that could yieldcabochons and faceted stones. Rarest of all are attrac-tive specimen-quality crystals (see, e.g., figure 11).Most of the production was sold in Antananarivo inDecember 2002–February 2003, to European andAmerican dealers. Additional smaller, lower-qualityparcels were sold until September 2003. It is not clearhow much of this material came from the originalfind, and how much resulted from later mining.African, Japanese, Thai, and Indian traders alsobought the material directly in Madagascar orthrough local buyers. The largest transparent facetedpezzottaite of good quality known to the authorsweighs approximately 11.31 ct, and the largest good-quality cat’s-eye cabochon is 17.36 ct (see figure 12).Examples of other large pezzottaite gems seen at theGIA and GRS laboratories include faceted stonesweighing 5.27, 6.26, and 10.07 ct, and cat’s-eye cabo-chons weighing 7.98, 8.78, and 8.85 ct.
As of November 2003, Dr. Pezzotta reportedthere was practically no material available in themine area, and only small, lower-quality materialcould be found in Antsirabe and Antananarivo.
MATERIALS AND METHODSThe samples and analytical techniques used in thisstudy are summarized in table 1. Gemologicalproperties were obtained on 19 samples (see, e.g.,figure 1) using standard gemological instruments.
PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003 289
Figure 11. Attractive crystals of pezzottaite arerelatively rare, particularly on matrix. This speci-men measures 4.7 × 4.2 cm, and consists of brightpink pezzottaite, feldspar, smoky quartz, andblack tourmaline. GIA Collection no. 30113;photo by R. Appiani.
Figure 10. Fine crystals of spodumene, some bicol-ored in violet and yellow as shown here (24.9 cmlong), were recovered from the same pocket withpezzottaite (here, 4.9 × 3.1 cm). Courtesy of BrianCook (pezzottaite) and the Brooke Collection (spo-dumene); photo by Maha Tannous.
Internal features were examined with a gemologi-cal microscope, and absorption spectra wereobserved with a desk-model Beck prism spectro-scope. The documentation of additional samplesexamined by one of the authors (AP) will be pub-lished in a separate article; these results are in gen-eral agreement with those obtained at GIA.
To test for color stability, one sample was sawninto four parts. Heating and irradiation experimentswere conducted on three parts, while the other onewas retained as a control.
Quantitative chemical analysis by electronmicroprobe was performed on 11 pezzottaite sam-ples from which a total of 49 point analyses wereobtained. Nearly all of these analyses were done onfragments that were donated for research; two of thegemstones mentioned above that could be borrowedlong enough for chemical analysis also were ana-lyzed. Analyses were obtained using an ARL-SEMQelectron microprobe with 15 kV (for sodium) and 25kV accelerating voltages, 15 nA beam current, and 3mm beam diameter. The measurements were cali-brated with natural mineral and synthetic com-pound standards, and a ZAF correction procedurewas applied to the data. Morphological observationsof eight etched crystals also were performed.
Quantitative chemical analysis by laser abla-tion–inductively coupled plasma–mass spectrome-try (LA-ICP-MS) using a GeoLas 193 nm excimerlaser in combination with an Elan 6100 ICP-DRC
(dynamic reaction cell) mass spectrometer was car-ried out on one sample. The water content of twoadditional fragments was determined by measuringthe weight lost after heating them to 900°C. In addi-tion, the lithium contents were measured in twoother samples using an ARL 3520 AES inductivelycoupled plasma spectrometer.
Visible–near infrared spectra of pezzottaite and,for comparison, morganite and red beryl wereobtained using a custom-built diode array spectrom-eter consisting of a tungsten-halogen source coupledto a microscope spectrometer. The detector, a 1025element silicon diode array, was attached to a grat-ing optical spectrograph that received its signalthrough fiber optics from the microscope.
Infrared spectra of pezzottaite, morganite, andred beryl were collected with a SensIR DuraScopediamond window attenuated total reflectance (ATR)apparatus. Spectra were obtained from a 0.1-mm-diameter area of finely powdered sample and werereferenced to a blank window. These spectra closelyresemble those obtained with the use of KBr pellets,but are considerably more convenient to obtain.
Raman spectra of pezzottaite, aquamarine, mor-ganite, and red beryl were obtained with aRenishaw 1000 microRaman system. Spectra weregenerated with both a 514.5 nm argon-ion laser anda 782 nm laser diode. Similar spectra were obtainedwith both lasers, although less fluorescence wasexcited with the 782 nm laser. In no case, however,did the fluorescence hinder the measurement. Allspectra were obtained with a depolarizer positionedbefore the samples, in orientations both parallel andperpendicular to the c-axis. The sloping baseline inthe spectra was corrected using a built-in Gramssoftware function and normalized to the intensityof the ~684 cm-1 band.
Powder X-ray diffraction analyses of pezzottaiteand aquamarine were performed using a Scintaginstrument with an accelerating voltage of 35 kVand a beam current of 15 mA.
RESULTSCrystal Forms and Visual Appearance. Rough. Thepezzottaite crystals showed hexagonal symmetryand consisted of three main forms in variousdegrees of prominence: the pinacoid c {0001}, pyra-mid d {101–2}, and prism m {101–0} (figures 13 and 14).Although not present on our samples, small second-order pyramid and prism faces have been document-ed (Warin and Jacques, 2003). The samples showed
290 PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003
Figure 12. At 17.36 ct, this is the largest good-qualitycat’s-eye pezzottaite known to the authors. Courtesyof K&K International; photo by Maha Tannous.
PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003 291
TABLE 1. Samples of pezzottaite from Ambatovita, Madagascar, and other samplesanalyzed for this study.
Samples (pezzottaite unlessotherwise noted) Method of analysis Analyzed at:1 Notes
19 total, all purplish pink: Gemological characterization: R.I., S.G.2, GIA Two of the faceted stones10 faceted (0.89–6.26 ct) optic character, Chelsea filter reaction, (1.18 and 3.42 ct) were 7 cabochons (1.16–8.78 ct) UV fluorescence, absorption spectros- analyzed by electron micro-1 crystal (7.5 grams) copy, and microscopic examination probe at UNO.1 partially polished fragment (1.64 ct)4 purplish pink fragments (~2 mm), Color stability: One fragment held as Caltech Another piece of this samplederived from one crystal a control, one fragment irradiated but was also used for ATR IR
not heated, and two fragments heated spectroscopy. in air for 2 hours at different temper-atures: 250, 350, and 450°C. The 450°C sample was irradiated by 6Mrads of Cs-137 gamma rays, at 0.9Mrads per day
11 total, all purplish pink: Chemical analysis by electron UNO R.I. values were obtained at3 faceted (0.31, 1.18, and 3.42 ct) microprobe (total of 49 point UNO on the fragment with the6 polished fragments analyses) lowest Cs content. Polished1 polished plate (5.08 mm thick) plate also used for Vis-NIR 1 sliced crystal spectroscopy.
1 purplish pink fragment (~1 ct) Laser ablation–inductively ETH R.I. values and density (mea-coupled plasma–mass spectrometry sured by a Berman balance) (LA-ICP-MS) were obtained at GRS
Swisslab on this fragment, which had the highest Cs con-tent measured in this study.
1 purplish pink polished plate Visible-NIR spectroscopy3 Caltech Spectrum reported in Simmons(5.08 mm thick) et al. (2003).1 polished plate of pale pink Same sample was used formorganite, Brazil (1.97 g) Raman spectroscopy.1 polished plate of red beryl, Same sample was used forWah Wah Mts., Utah (0.38 g) Raman spectroscopy.1 polished plate of red beryl,Thomas Range, Utah
1 purplish pink powdered Attenuated total reflectance Caltechsample (ATR) IR spectroscopy1 powdered sample of pale From the same sample used pink morganite, Brazil for Vis-NIR and Raman. 1 powdered sample of redberyl, Wah Wah Mts., Utah
2 purplish pink polished plates Raman spectroscopy Caltech(6.9 and 0.4 mg)1 polished plate of pale pinkmorganite, Brazil (1.97 g)1 polished plate of red beryl,Wah Wah Mts., Utah (0.38 g)1 polished plate of pale blue aqua-marine, Minas Gerais, Brazil (0.38 g)
1 purplish pink powdered sample X-ray powder diffraction UNO(~80 mg)1 powdered sample of pale blue aqua-marine, Erongo Mts., Namibia (~80 mg)
1 Abbreviations: Caltech = California Institute of Technology, ETH = Eidgenössische Technische Hochschule,UNO = University of New Orleans.2 Note that the S.G. value of the partially polished fragment was not included, owing to inconsistent results thatwere probably due to trapped air bubbles within surface irregularities.3 Spectra were obtained at approximately 1.5 nm resolution.
varying degrees of etching and corrosion, with resid-ual light-colored spongy masses present along theprism areas of some crystals. The underlying vitre-ous surfaces (seen where the spongy masses brokeoff or were entirely removed by corrosion) common-ly showed a series of microsteps in reflected light.Warin and Jacques (2003) interpreted these surfacesas resulting from the original crystal growth, ratherthan corrosion, but this interpretation is not sup-ported by our observations.
Besides the spongy corroded areas mentionedabove, all of the samples were purplish pink.Dichroism (described below) was not readily seenwhen observing our samples with the unaided eye(due to their relatively small size), but it is quite obvi-ous in larger samples (see, e.g., figures 14 and 15).
Polished. The polished stones we examined weretransparent to translucent, deep purplish pink, andevenly colored with no eye-visible color zoning(again, see figure 1). They displayed moderate tostrong dichroism: pinkish orange to pink-orangewhen viewed down the c-axis (w ray) and purplishpink to pink-purple perpendicular to the c-axis. Insome larger stones, depending on how they were cutrelative to the optic axis, the pink-orange pleochro-ism could be observed with the unaided eye.
GEMOLOGICAL CHARACTERISTICSPhysical and Optical Properties. The results ofgemological testing are summarized in table 2. Therefractive indices of the faceted stones generally
were no = 1.615–1.619 and ne = 1.607–1.610 (birefrin-gence 0.008–0.009), with the cabochons giving a spotreading of 1.61. As expected, the R.I. values of thefragment with the lowest Cs content (as analyzed byelectron microprobe) were somewhat lower thanthose obtained on the faceted stones, at no = 1.612and ne = 1.601. The samples yielded S.G. values of3.09–3.14. Although the S.G. values of our cabo-chons were not significantly influenced by the pres-ence of growth tubes, it would not be surprising toobtain a lower S.G. for samples containing largenumbers of them.
Pezzottaite is uniaxial negative, although sever-al samples were anomalously biaxial in their opticfigures, where the “brushes” were seen to cometogether and move apart slightly as a stone wasrotated around the optic axis. This effect was mostlikely due to strain, which was visible in some ofthe samples when viewed parallel to the optic axisbetween crossed polarizing filters. The samplesappeared faint orangy pink to pink-orange whenviewed through a Chelsea color filter with a dif-fused light source; too strong a light source over-powered the reaction and the stone appeared yel-low-green. All samples were inert to long- and
292 PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003
Figure 13. This idealized diagram shows the promi-nent faces of pezzottaite crystals, which consist of c{0001}, d {101
–2}, and m {101
–0} forms. The drawing
was generated using Kristall2000 software that wasdeveloped by Klaus Schilling.
short-wave UV radiation, although one sample didfluoresce a weak chalky yellow in the fractures,which indicated the presence of a filling material.
All samples displayed oriented absorption spec-tra with the desk-model spectroscope. When thespectrum was viewed with the stone parallel to thec-axis (down the optic axis), diffuse lines were visi-ble at 465 and 477 nm, along with a band at approx-
imately 485–500 nm. In some samples, these threefeatures converged into a single broad absorptionband. When viewed perpendicular to the c-axis,bands at approximately 485–500 and 550–580 nmwere visible. In all other orientations, combina-tions of the above-listed absorption features wereobserved with varying intensities.
Internal Features. All of the 19 samples we exam-ined contained numerous fine growth tubes (figure16) oriented parallel to the c-axis. When abundant,these tubes are responsible for the chatoyancy seenin cat’s-eye pezzottaite. Some of these tubes hadnegative crystal “flare outs” along their length (fig-ure 17). These features are indicative of an abruptchange in the growth environment (i.e., slowergrowth; J. I. Koivula, pers. comm., 2003). The inten-sity of strain seen between crossed polarizersseemed to increase with the number of growthtubes. Fractures, “fingerprints,” and fluid inclusionsalso were present in all of the samples. Some of thesamples contained transparent, near-colorless, low-
PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003 293
Figure 16. All of the pezzottaite samples examinedcontained numerous fine growth tubes oriented par-allel to the c-axis. Some of the tubes shown here canbe seen emanating from included crystals (identifiedas tourmaline). Photomicrograph by John I. Koivula;magnified 15×.
TABLE 2. Properties of pezzottaite from Madagascar.
Property Description
Color • Purplish pink• Moderate to strong dichroism: pinkishorange to pink-orange (w ray) and pur-plish pink to pink-purple (e ray)
Diaphaneity Transparent to translucentR.I. • no = 1.615–1.619, ne = 1.607–1.610
(faceted stones); spot reading = 1.61 (cabochons)
• no = 1.612, ne = 1.601 for polishedfrag-ment with the lowest Cs content
• no = 1.620, ne = 1.611 for polishedfrag-ment with the highest Cs content
Birefringence 0.008–0.009Optic character Uniaxial negative, with several samples
showing anomalously biaxial opticsS.G. Typically 3.09–3.11, with faceted
stones—3.09–3.14, cat’s-eye cabochons—3.09–3.11, and crystal—3.09
Chelsea filter reaction Faint orangy pink to pink-orangeUV fluorescence Inert to long- and short-wave UV radiationa
Spectroscope spectra • Viewed parallel to optic axis: diffuse linesat 465 and 477 nm plus a band at 485–500 nm
• Viewed perpendicular to optic axis: bands at approximately 485–500 nm and 550–580 nm
Internal features Growth tubes, fractures, “fingerprints,” andfluid inclusions (and, rarely, noticeable two-phase inclusions); transparent, near-color-less, low-relief birefringent crystals; trans-parent, grayish green tourmalinecrystals; strain; rare color zoning
aWeak chalky yellow fluorescence in fractures of one sample sug-gested the presence of a filling.
Figure 15. Pleochro-ism in purplish pinkand pink-orange isclearly visible in thislarge piece of pezzot-taite, which measures4.9 × 3.1 cm. Photosby Maha Tannous.
relief birefringent crystals, which could not be iden-tified due to their small size and their position with-in the stones. A few samples contained transparent,grayish green, birefringent crystals that were identi-fied (by Raman analysis) as tourmaline (figure 18).One sample contained obvious two-phase fluid andgas inclusions. This sample also showed color zon-ing, but we observed no notable growth or colorzoning in the other samples.
Although the dealers who supplied the sampleswere not aware of any clarity enhancement, many of
the samples showed fractures of low relief, some ofwhich contained air bubbles that contracted whenexposed to a thermal reaction tester, indicating thepresence of a filling substance. Subsequently, welearned from Dudley Blauwet (pers. comm., 2003)that all of the pezzottaite rough he has purchasedwas oiled by the local Madagascar dealers. He report-ed that the oiling enables them to see into the rougheasier (because of the slight etching on most of thecrystal surfaces). While this may be true, pezzottaite,like many gem materials that tend to be fractured,can easily be clarity enhanced with oil or resin toimprove its appearance.
Heating and Irradiation Experiments. When heated at450°C, a purplish pink fragment of pezzottaite suf-fered a near-total loss of color, although no effect wasnoted at lower temperatures. The sample regainednearly all of its purplish pink color on irradiationwith gamma rays.
CHEMISTRY The highest Cs content was obtained on the sampleanalyzed by LA-ICP-MS: 18.23 wt.% Cs2O (table 3).The 11 samples analyzed by electron microprobehad cesium contents ranging from 11.23 to 15.13wt.% Cs2O (again, see table 3). A similar range alsowas obtained from four samples analyzed by elec-tron microprobe at the University of Manitoba byone of the authors (FCH; unpublished data).Calculations of Cs ions per formula unit yielded0.504 to 0.833.
Minor amounts of Rb and Na were also mea-sured in all samples. Manganese, the chromophoricelement in pink beryl, ranged up to 0.19 wt.% MnO,with the average being 0.11 wt.%. Mg, K, Sc, Ti, andFe were either below the detection limit or presentonly in trace amounts in some samples. Many otherelements were looked for, but not detected (see foot-note to table 3). To view all 49 electron microprobeanalyses, see table 1 in the Gems & Gemology DataDepository at www.gia.edu/gemsandgemology.
SPECTROSCOPYVis-NIR. With the beam oriented down the c-axis(or E^c), the Vis-NIR absorption spectrum of pur-plish pink pezzottaite is dominated by a band cen-tered at 494 nm, with a distinct shoulder at 476nm (figure 19). A second absorption band at 563nm appears in the same polarization, along with a
294 PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003
Figure 18. Raman analysis identified the grayish greeninclusions in this pezzottaite as tourmaline.Photomicrograph by John I. Koivula; magnified 15×.
Figure 17. Negative crystal “flare outs” were some-times seen along growth tubes in pezzottaite. Note thebipyrimidal form of these minute negative crystals.Photomicrograph by John I. Koivula; magnified 40×.
broad band that is centered at 820 nm in the near-infrared region. The maximum transmissionoccurs near 630 nm (orange-red) and in the low-400 nm range (deep violet), which provides thepinkish orange pleochroic color in this direction.When the beam is oriented perpendicular to the c-axis (i.e., E||c), the spectrum is dominated by aband centered at 572 nm. A weak absorption bandin the near-infrared region is also present at 820nm. Transmission in the other regions of the spec-trum combine to produce a purplish pink color forlight polarized in this direction.
Other Methods. The infrared and Raman spectra, aswell as X-ray diffraction patterns, of pezzottaite arebriefly described here, and can be viewed in theGems & Gemology Data Depository.
The ATR infrared spectrum generally resembledthat of beryl (i.e., morganite and red beryl), withsome shifts in wavenumbers and relative intensi-ties. The most prominent differences were evidentat higher wavenumbers; the absorption in the1400–1100 cm-1 region occurs at a much higherwavenumber in pezzottaite than in the other min-erals. The absorption in the region from about 1200to 800 cm-1 correlates to various Si–O and Be–Obands (Hofmeister et al., 1987).
The characteristic feature of the pezzottaiteRaman spectrum was the 1100 cm-1 band, whichhas not been observed in the spectra of other pinkor red beryls or pale aquamarine. This band waspresent in Raman spectra obtained from both the(0001) and the ||c face.
In the X-ray diffraction pattern of pezzottaite,some important peaks typical of beryl were absentor very much weaker in intensity. Indexing ofthese patterns indicated that the affected peakscorresponded principally to h00 type reflections.
DISCUSSION Cs in Beryl and Pezzottaite. Alkali beryls contain-ing Cs have been informally referred to in the litera-ture by some mineralogists as vorobievite—com-monly for the pink Li-Cs variety—and rosterite—typically for the near-colorless Na-Li type (Beus,1966; see also Zambonini and Caglioti, 1928;Rossovskiy, 1981). Morganite is the most commonterm used to refer to gem-quality pink beryl; thiscolor variety commonly contains Cs. The amountof Cs (and other alkalis) that is incorporated intoberyl has been correlated to the local abundance of
TABLE 3. Chemical analyses of pezzottaite from Ambato-vita, Madagascar, by LA-ICP-MS and electron microprobe.a
LA-ICP-MS Electron microprobebLowest Csc Highest Csd Averagee
a Abbreviations: nd = not determined, bdl = below detection limit.b For the electron microprobe data, values for lithium (by ICP) andwater (by LOI) were measured for two samples (Simmons et al.,2003). The measured Li2O value was then used as the best approx-imation for calculating beryllium (Be+Li = 3) and the other ions performula unit for all samples analyzed. The H2O value is consideredanomalously high due to the presence of microscopic tubules that are filled with an aqueous fluid (consistent with the granitic peg-matite environment). The low analytical totals (as calculated withoutwater) are attributed to the abundance of these fluid-rich inclusions.Crystal structure refinement of a sample by FCH indicated only 0.28wt.% H2O in the channels. The following were analyzed by electronmicroprobe but were below the detection limits (shown in wt.%): MgO (0.01), TiO2 (0.002), FeO (0.02), CaO (0.07). In addition, the following were checked for (scanned), but not detected: Cr2O3(0.03), Bi2O3 (0.03), V2O3 (0.03), PbO (0.01), ZnO (0.08), BaO (0.03), Cl (0.04), F (0.05).c Fragment of a purplish pink sample.d Polished plate used for Vis-NIR spectroscopy reported in Simmons et al. (2003).e Average of 49 analyses of 11 samples.f Ca is assumed present in the octahedral site, but it may occur elsewhere in the beryl structure.
PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003 295
such elements in the growth environment ( `Cerny,1975). The alkalis reside in the channel site, and areaccommodated into the beryl structure throughcoupled substitutions (see box A).
The concentrations of the alkali elements Cs, Rb,and Li have been measured in beryl from a variety ofgeologic environments (Staatz et al., 1965). Thegreatest amounts of these elements have been foundin beryl from the inner zones of granitic pegmatites,since this is where such elements—which areincompatible with common pegmatite mineralssuch as quartz and feldspar—are concentrated duringcrystallization (Beus, 1966; `Cerny, 1975; Zagorsky etal., 1999; `Cerny et al., 2003). Geochemically, peg-matitic beryl samples with the highest Cs contentsgenerally also show low values of Na/Li. Typically,high-Cs beryls range up to 4 wt.% Cs2O, althoughthere are a few notable exceptions (table 4). At 11.3wt.% Cs2O, it is probable that the Madagascar sam-ple studied by Evans and Mrose (1966) was pezzot-taite. Furthermore, `Cerny (1972) reported thatminute crystals of late hydrothermal beryl from theTanco pegmatite had R.I. values exceeding thosereported by Evans and Mrose (1966). Although theywere not chemically analyzed, and specific R.I. val-ues were not reported, these crystals also may havecorresponded to pezzottaite.
According to Dr. H. A. Hänni (pers. comm.,2003), the Afghanistan sample with 9.70 wt.%Cs2O reported by Hänni and Krzemnicki (2003) isactually pezzottaite (based on its unit-cell dimen-sions). Therefore, this sample has the lowest Cscontent found in pezzottaite so far. The highest con-centration of Cs measured in pezzottaite wasobtained on samples from Madagascar byAbduriyim and Kitawaki (2003), who measured19.76–21.33 wt.% Cs2O by X-ray fluorescence spec-troscopy and up to 23.37 wt.% Cs2O by LA-ICP-MS.
Early studies showed a strong relationshipbetween alkali content and refractive indices (as wellas specific gravity) in beryl from Madagascar (see, e.g.,Ford, 1910; Lacroix and Rengade, 1911; Lacroix,1912). More recently, the relationship between com-position and optical/physical properties was docu-mented in beryl from several localities (reviewed byDeer et al., 1997).
`Cerny and Hawthorne (1976) pointed out that fac-tors in addition to alkali content can influence R.I.values, so a simple mathematical relationshipbetween them is impossible to generalize. This isshown by the high R.I. values reported for Cs-richberyl in the literature (table 4). (Note that the highproperties of the Arizona sample were attributed toenriched contents of both Fe [4.69 wt.% “oxides”]and Cs.) Since Cs is a relatively heavy element, sig-nificant concentrations of it would be expected to
296 PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003
Figure 19. Representative polarized Vis-NIR spectra areshown here for a 5.08-mm-thick slice of pezzottaite(top), a 10.79-mm-thick slice of morganite from Brazil(middle), and a 0.72-mm-thick slice of red beryl fromthe Thomas Range, Utah (bottom). The pezzottaite isdominated by bands centered at 494 and 563 nmwhen the beam is oriented down the c-axis (or E⊥c),and a band centered at 572 nm when the beam is ori-ented perpendicular to the c-axis (i.e., E||c). The differ-ence in these absorptions accounts for pezzottaite’smoderate dichroism in purplish pink and pink-orange.The morganite shows similar absorptions that areshifted somewhat and not as intense. This relativelythick sample also shows overtones of the water absorp-tion near 957 nm. The red beryl shows intense absorp-tions centered at 527 and 559 nm; the minimal differ-ence between the two red beryl spectra corresponds tothe lack of noticeable pleochroism in this material.
PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003 297
raise S.G. values as well. By comparison to the val-ues reported in table 4, the theoretical specific gravi-ty of beryl that has been calculated using X-raydiffraction measurements is 2.661 (Deer et al., 1997).
Origin of Color. An early study by Wood andNassau (1968) attributed the color of morganite toMn2+ and that of red beryl to Mn3+, an assignmentthat Schmetzer et al. (1974) reiterated. Platonov etal. (1989) re-examined the spectra of manganese-containing beryls and concluded that Mn3+ is
responsible for the color of both red beryl and mor-ganite. Likewise, Solntsev and Bukin (1997) con-cluded that the pink morganite from Mozambiqueis colored by Mn3+ (specifically, d-d transitions inMn3+ in the aluminum site).
In the visible range of the spectrum, the mostsignificant difference between pezzottaite and mor-ganite is the overall greater absorption intensity ofpezzottaite, corresponding to its deeper color (figure19). In fact, the spectra of the pezzottaites we stud-ied have 5.3 times greater absorbance per centime-
How is Cs (a relatively large ion) incorporated intothe beryl structure? This mechanism was studied byHawthorne and `Cerny (1977), and examined in moredetail (particularly for lithium) by Sherriff et al.(1991). A useful review of beryl crystal chemistry wasprovided by `Cerny (2002), and the current consensuswas summarized by `Cerny et al. (2003, p. 1006):“lithium substitutes for Be2+ in the tetrahedra linkingthe six-membered Si6O18 rings, with the charge bal-ance assured by Na+ and Cs+ in the channels passingthrough the centers of the vertically stacked rings.Sodium is accommodated in the centers of the indi-vidual rings of [silicate] tetrahedra, but the larger Cs+
is located halfway between these centers” along thec-axis. The increase in charge caused by the incorpo-ration of alkalis into the channel is compensated byreplacement of Be by Li at a tetrahedral site.
Pezzottaite, ideally Cs(Be2Li)Al2Si6O18, is not
isostructural with beryl, ideally Be3Al2Si6O18.However, the arrangement of the atoms in eachstructure is very similar, so the structures can beconsidered as very closely related. In beryl, the threeBe atoms indicated in the formula occupy threesymmetrically related positions. In pezzottaite,these three positions are occupied in an orderedfashion by Be2Li rather than Be3 (as in beryl). Thus,these positions in pezzottaite are no longer relatedby symmetry, so the symmetry of the mineral mustchange. As a result, pezzottaite differs from othermembers of the beryl group by having a differentsymmetry (rhombohedral rather than hexagonal)and a larger unit cell.
Further studies are needed to characterize thedividing line between pezzottaite and Cs-rich beryl,which is influenced by factors such as composition(Be-Li ordering), symmetry, and unit cell dimensions.
BOX A: STRUCTURE OF BERYL AND PEZZOTTAITE,AND DEFINITION OF THE MINERAL
TABLE 4. Properties of high-Cs beryls and pezzottaite.a
Wt.%Description Cs2O no ne S.G. Reference
“Bluish” beryl from Arizona 6.68 1.608 1.599 2.92 Schaller et al. (1962)Beryl from Tanco, Bernic Lake 7.16 nr nr nr `Cerny (1972)district, CanadaMorganite from Madagascar 7.52 nr nr nr Cabell and Smales
(1957)Pezzottaite from Afghanistan 9.70 1.604 1.598 2.91 Hänni and
Krzemnicki (2003)Pezzottaite from Madagascar 11.23 1.612 1.601 nd This study (lowest
Cs)“Cesium beryl” from Antsirabe 11.3 1.608 1.601 3.01 Evans and Mrosearea, Madagascar (1966)Pezzottaite from Madagascar 18.23 1.620 1.611 3.09 This study (highest
Cs)
a nr = not reported, nd = not determined
ter than the one for morganite presented inPlatonov et al. (1989). The spectrum of our morgan-ite was similar to the Platonov et al. data, but was afactor of 10 less intense. While the spectrum of redberyl (again, see figure 19) shows absorption bandsat generally the same wavelengths, their intensitydistribution is distinct. The stability of the color ofred beryl at high temperatures suggests that itscolor is not due to irradiation, and thus leads to thespeculation that manganese in red beryl was origi-nally present as Mn3+ during crystallization. In con-trast, the instability of the manganese-derived colorin morganite and pezzottaite at modest tempera-tures leads to the speculation that these gems origi-nally formed containing Mn2+, but became pink as aresult of exposure to natural ionizing radiation aftercrystallization. The presence of such radiation inthe Sakavalana pegmatite is demonstrated by thecommon association of smoky quartz with morgan-ite and pezzottaite in the gem-bearing cavities; theradiation may be derived from radioactive traceminerals or the isotope 40K in K-feldspars.
Previous studies (Nassau, 1984) have shown thatsome pink beryls, when decolorized by heat, regaintheir color upon irradiation with X-rays or gamma rays.This is consistent with our results for pezzottaite, andthe sensitivity to heating and irradiation supports ourproposal that the color of pezzottaite is caused by radia-tion-induced color centers involving Mn3+.
It may be tempting to correlate the pink color ofpezzottaite (as well as high-Cs morganite) withcesium, but this element has been ruled out as achromophore in beryl (Ristic and Eichoff, 1955;Sinkankas, 1981). Although Cs-rich beryls do com-monly have a pink color, this is probably due to thesimple fact that Cs and Mn follow each other geo-chemically toward the end of crystallization ingranitic pegmatites (see `Cerny et al., 1985). Thus, itis likely that both Cs and Mn will be present during
the crystallization of beryl (or pezzottaite) in peg-matite pockets.
Identification. The identification of pezzottaiteshould be straightforward once the gemologist isfamiliar with its properties and a flat surface for anR.I. reading is available. The optical and physicalproperties of pezzottaite are distinct, particularlywhen compared to the varieties of beryl that itresembles—morganite and red beryl.
All of the samples of pezzottaite we examinedwere purplish pink. Although they varied some-what in tone and saturation, their color was neverthe same as typical red beryl from Utah or its syn-thetic counterpart. This fact alone should preventany confusion with these materials. In addition, theR.I. and S.G. values of pezzottaite are significantlyhigher than those of red beryl, for which the typicalpublished values are no = 1.568–1.572, ne =1.564–1.569, and S.G. = 2.66–2.70 for natural mate-rial (Shigley and Foord, 1984), and no = 1.576–1.580,ne = 1.569–1.573, and S.G. = 2.67–2.70 for the syn-thetic counterpart (Shigley et al., 2001). So far, pez-zottaite has not been synthesized in the laboratory;the highest cesium content in synthetic berylknown to these authors is 2.39 wt.% Cs2O(Shatskiy et al., 1981).
Some morganite may have a color similar to pez-zottaite. However, the two can easily be separated gemologically by the distinctly lower R.I. and S.G.values of morganite (typically with R.I. values of no =1.572–1.592 and ne = 1.578–1.600, and S.G. =2.71–2.90; Arem, 1987). Compared to pezzottaite,even high-Cs beryls have lower values (e.g., general-ized by Schaller et al., 1962, as no = 1.599, ne = 1.590,and S.G. ª 2.86).
The color of pezzottaite is almost the same assome tourmaline (figure 20, left). Pink tourmalinepresents a more significant identification challenge
298 PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003
Figure 20. As seen in the photo on the left, cat’s-eye pink tourmaline (left stone, 9.59 ct) is very similar in appear-ance to pezzottaite (right stone, 8.78 ct). However, they can be readily separated by their pleochroism, which isstronger in pezzottaite. This effect can be seen in the center and right photos, which were taken through a polariz-ing filter at different orientations. The pezzottaite is courtesy of William Pinch and the tourmaline is from the GIACollection (no. 5738); photos by Maha Tannous.
than do other beryls, particularly since many of thefinest examples of pezzottaite are cat’s-eyes. Whenencountering a pink cat’s-eye, the gemologist islikely to think of tourmaline, since there are fewother materials of this color that show chatoyancy.The refractive index of a pezzottaite cabochon mea-sured by the spot method will read ~ 1.61, which isvery close to the 1.62 that one would expect fromtourmaline. Microscopically, the inclusions in pez-zottaite also resemble those in tourmaline, beingmostly parallel acicular tubes and planes of liquidinclusions. Even if the equipment is available formeasuring the S.G. hydrostatically, the value willfall into the range of tourmaline (typically 3.01–3.06for pink to red material [Arem, 1987]).
So, how does one separate these two materials?A faceted stone should not pose a problem, since anaccurate refractive index measurement will clearlyseparate the two gem minerals. If a gemologist isgood at taking spot R.I. readings, a value of 1.61should be enough to warrant closer inspection. Thepleochroism in pezzottaite is more pronouncedthan in pink tourmaline (figure 20, center and right).The purplish pink and pink-orange pleochroic direc-tions also are different for these two materials, andtheir visibility in a sample will depend on the orien-tation of the polarizer. The best way to check thepleochroism would be to compare an unknownsample with a known piece of pink tourmaline. Adesktop spectroscope can also be very useful, sincethe two minerals have distinctly different spectra;pink to red tourmaline typically shows narrow linesat 450 and 458 nm, and a broad region of absorptionthat is centered at ~ 525 nm (Webster, 1994). Ofcourse, the best way to avoid misidentifying a pez-zottaite cat’s-eye as tourmaline is to be aware of theexistence of this new mineral.
Additional Comparisons to Morganite and UtahRed Beryl. The chemical compositions of pezzot-taite, morganite, and red beryl are distinct. Besidesbeing virtually anhydrous, red beryl contains muchhigher amounts of Fe, Ti, Mn, and other trace ele-ments (Shigley et al., 2003; see also table 2 in theG&G Data Depository). Morganite shows ranges ofFe, Ti, and Mn similar to those of pezzottaite,although the latter may contain significantly moreMn (typically <0.05 wt.% in morganite, vs. an aver-age for this study of 0.11 wt.% MnO in pezzottaite).
Differences between pezzottaite and beryl in theinfrared and Raman spectra, as well as in X-raydiffraction patterns (again, see Data Depository),
appear distinctive and therefore provide additionalcriteria for separating the two minerals. What is notknown at this time is how these properties varywith Cs content, and if beryls with intermediatelevels of Cs can be distinguished from pezzottaiteon the basis of these techniques.
Although these authors do not have first-handexperience with mounting pezzottaite in jewelry,the material is expected to behave similar to beryl innormal manufacturing, wear, and care conditions.
CONCLUSIONDue to its high Cs and Li contents and a structurethat shows differences from beryl, pezzottaite hasnow been recognized as the fourth mineral of theberyl group. Unlike most new mineral species iden-tified in recent years, pezzottaite was found as rela-tively large, well-formed crystals and fragments, andsome of the rough proved suitable for cutting attrac-tive gems (figure 21). The enriched Cs content gives
PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003 299
Figure 21. Although considered a collector’s stonedue to its rarity, pezzottaite has been set into someattractive jewelry. This gold ring, containing a 5.25 ctpezzottaite, was created by Francis Bonnet ofPolychrome France Co., Chambray-lès-Tours.Courtesy of Laurent Thomas.
REFERENCESAbduriyim A., Kitawaki H. (2003) Analysis on Cs pink “beryl”
using a laser ablation system with inductively coupled plasmamass spectrometer (LA-ICP-MS). Gemmology, Vol. 34, No.411, pp. 24–26 (in Japanese, with insert of English translation).
Arem J.E. (1987) Color Encyclopedia of Gemstones. VanNostrand Reinhold Co., New York.
Armbruster T., Libowitzky E., Diamond L., Auernhammer M.,Bauerhansl P., Hoffmann C., Irran E., Kurka A., RosenstinglH. (1995) Crystal chemistry and optics of bazzite fromFurkabasistunnel (Switzerland). Mineralogy and Petrology,Vol. 52, pp. 113–126.
Aurisicchio C., Fioravanti G., Grubessi O., Zanazzi P.F. (1988)Reappraisal of the crystal chemistry of beryl. AmericanMineralogist, Vol. 73, pp. 826–837.
Beus A.A. (1966) Geochemistry of Beryllium and Genetic Types ofBeryllium Deposits. Ed. by L.R. Page, and transl. by F. Lachmanand R. K. Harrison. W.H. Freeman and Co., San Francisco.
Cabell M.J., Smales A.A. (1957) The determination of rubidiumand caesium in rocks, minerals and meteorites by neutron-activation analysis. The Analyst, Vol. 82, pp. 390–406.
`Cerny P. (1972) The Tanco pegmatite at Bernic Lake, Manitoba.VII. Secondary minerals from the spodumene-rich zones.Canadian Mineralogist, Vol. 11, pp. 714–726.
`Cerny P. (1975) Alkali variations in pegmatitic beryls and theirpetrogenetic implications. Neues Jahrbuch fur Mineralogie,Abhandlungen, Vol. 123, pp. 198–212.
`Cerny P. (1991) Rare-element granitic pegmatites. Part I:Anatomy and internal evolution of pegmatite deposits.Geoscience Canada, Vol. 18, pp. 49–67.
`Cerny P. (2002) Mineralogy of beryllium in granitic pegmatites.In E.S. Grew, Ed., Beryllium: Mineralogy, Petrology, andGeochemistry, Reviews in Mineralogy and Geochemistry,Mineralogical Society of America, Washington D.C., Vol. 50,pp. 405–444.
`Cerny P., Hawthorne F.C. (1976) Refractive indices versus alkalicontents in beryl: General limitations and applications to somepegmatitic types. Canadian Mineralogist, Vol. 14, pp. 491–497.
`Cerny P., Meintzer R.E., Anderson A.J. (1985) Extreme fractionationin rare-element granitic pegmatites: Selected examples of dataand mechanisms. Canadian Mineralogist, Vol. 23, pp. 381–421.
rise to some of its properties, which permit astraightforward separation from beryl and otherpink gem minerals, except for cabochons of pinktourmaline, which can be identified by their dichro-ism and absorption spectra.
Significant amounts of gem-quality pezzottaitethus far have been found only at the Sakavalanapegmatite in a remote area of central Madagascar.Although the pegmatite has been mined since the
1940s for tourmaline, pezzottaite was not recov-ered there until mid-November 2002. The mineral-ized area has been mostly mined out by artisanalmethods, so any future production of pezzottaitewill depend on the use of mechanized mining andsystematic exploration of the deposit. With therecognition of this material as a new gem mineral,it is also possible that it will be identified in othergranitic pegmatites.
300 PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003
ABOUT THE AUTHORS
Mr. Laurs ([email protected]) is editor of Gems & Gemology atGIA in Carlsbad, California. Dr. Simmons is professor of min-eralogy and university research professor, and Mr. Falster issenior research technologist, at the University of New Orleans,Louisiana. Dr. Rossman is professor of mineralogy at theCalifornia Institute of Technology, Pasadena. Ms. Quinn is staff gemologist, and Mr. McClure is director of IdentificationServices, at the GIA Gem Laboratory, Carlsbad. Dr. Peretti isdirector of the GRS Gemresearch Swisslab Ltd., Lucerne,Switzerland. Dr. Armbruster is professor of mineralogical crys-tallography at the University of Bern, Switzerland. Dr.Hawthorne is professor of mineralogy, and Mr. Cooper is lab-oratory technician, at the University of Manitoba, Winnipeg,Canada. Dr. Günther is professor for trace elements andmicroanalysis at the Laboratory for Inorganic Chemistry, ETHZurich, Switzerland. Dr. Grobéty is professor of mineralogy atthe University of Fribourg, Switzerland.
ACKNOWLEDGMENTS: The authors are grateful to Dr.Federico Pezzotta (Natural History Museum, Milan, Italy) forsupplying detailed information on the geology, mineralogy, and morphology of pezzottaite. The senior author is indebtedto Dr. Pezzotta for guiding his visit to the mine in July 2003.We thank the following gem dealers for loaning and/or donat-ing research samples: Mark Kaufman of Kaufman Enterprises,
San Diego, California; Dudley Blauwet of Dudley BlauwetGems, Louisville, Colorado; Laurent Thomas of PolychromeFrance Co., Chambray-lès-Tours, France; Denis Gravier andFabrice Danet of Le Mineral Brut, Saint-Jean-le-Vieux, France;Tom Cushman of Allerton Cushman & Co., Sun Valley, Idaho;Marc Jobin and Steve Jaquith of MJ3 Inc., New York; HerbObodda of H. Obodda, Short Hills, New Jersey; StuartWilensky of Stuart and Donna Wilensky Fine Minerals,Wurtsboro, New York; Irv Brown of Irv Brown Fine Minerals,Fallbrook, California; Edward Boehm of Joeb Enterprises,Solana Beach, California; K & K International, Falls Church,Virginia; William Pinch of Pittsford, New York; Rob Lavinsky ofThe Arkenstone, San Diego; Brian Cook of Nature’s Geometry,Graton, California; G.E.O. International Co. Ltd of Bangkok;and Papas Gem Co. Ltd., Bangkok. We thank John I. Koivulaof GIA, Carlsbad, for supplying samples and photomicro-graphs. Dr. Petr `Cerny of the University of Manitoba, Winnipeg, provided useful comments on the manuscript.Assistance with document scanning and computer-automat-ed translations was provided by Denise Breceda of the GIAGem Laboratory, Stuart Overlin of Gems & Gemology, andSheryl Elen of the Richard T. Liddicoat Library and InformationCenter in Carlsbad. Neil Barron and Ruth Patchick of theLiddicoat library obtained numerous publications via interli-brary loan. Portions of Zagorsky et al. (1999) were translatedby Inna Saphonova, Novosibirsk, Russia.
`Cerny P., Anderson A.J., Tomascak P.B., Chapman R. (2003)Geochemical and morphological features of beryl from theBikita granitic pegmatite, Zimbabwe. Canadian Mineralogist,Vol. 41, pp. 1003–1011.
Chantraine J. (1966) Etude géologique et prospection au1/100,000 des feuilles Janjina-Mandrosonoro (J-K, 51)[Geological investigation and prospection of the 1/100,000sheets Janjina-Mandrosonoro (J-K, 51)]. Travaux du BureauGéologique, No. 58, Service Géologique, Tananarive.
Deer W.A., Howie R.A., Zussman J. (1997) Rock-formingMinerals—Disilicates and Ring Silicates, Vol. 1B, 2nd ed. TheGeological Society, London, pp. 372–409.
Della Ventura G., Rossi P., Parodi G.C., Mottana A., RaudseppM., Prencipe M. (2000) Stoppaniite, (Fe,Al,Mg)4(Be6Si12O36) •(H2O)2(Na, ) a new mineral of the beryl group from Latium (Italy).European Journal of Mineralogy, Vol. 12, pp. 121–127.
Evans H.T., Mrose M.E. (1966) Crystal chemical studies ofcesium beryl. Program & Abstracts, Geological Society ofAmerica Annual Meeting, San Francisco, p. 63.
Ferraris G., Prencipe M., Rossi P. (1998) Stoppaniite, a new memberof the beryl group: Crystal structure and crystal-chemical impli-cations. European Journal of Mineralogy, Vol. 10, pp. 491–496.
Ford W.E. (1910) The effect of the presence of alkalies in berylupon its optical properties. American Journal of Science, Vol.30, pp. 128–130.
Hänni H.A., Krzemnicki M.S. (2003) Caesium-rich morganitefrom Afghanistan and Madagascar. Journal of Gemmology,Vol. 28, No. 7, pp. 417–429.
Hawthorne F.C., `Cerny P. (1977) The alkali-metal positions inCs-Li beryl. Canadian Mineralogist, Vol. 15, pp. 414–421.
Hawthorne F.C., Cooper M.A., Peretti A., Simmons W.B.,Armbruster T., Rossman G.R., Günther D., Laurs B.M.,Grobéty B. (2003a) Check-list for new mineral proposals:Pezzottaite. Proposal submitted to the InternationalMineralogical Association, 8 pp.
Hawthorne F.C., Cooper M.A., Simmons W.B., Falster A.U.,Laurs B.M., Armbruster T., Rossman G.R., Peretti A.,Günther D., Grobéty B. (in preparation) Pezzottaite,Cs(Be2Li)Al2Si6O18, a spectacular new mineral related to theberyl group, from Madagascar.
Hofmeister A.M., Hoering T.C., Virgo D. (1987) Vibrational spec-troscopy of beryllium aluminosilicates: Heat capacity calcula-tions from band assignments. Physics and Chemistry ofMinerals, Vol. 14, No. 3, pp. 205–224.
Lacroix A. (1912) Sur la continuité de la variation des propriétésphysiques des béryls de Madagascar, en relation avec leurcomposition chimique. Bulletin de la Société Française deMinéralogie, Vol. 33, pp. 200–208.
Lacroix A., Rengade (1911) Sur les propriétés optiques des bérylsroses de Madagascar. Bulletin de la Société Française deMinéralogie, Vol. 34, pp. 123–125.
Laurs B.M. (2003) Gem News International: Update on sapphire,pezzottaite, and other gems from Madagascar. Gems &Gemology, Vol. 39, No. 3, pp. 229–231.
Meagher E.P., Gibbs, G.V. (1977) The polymorphism ofcordierite; II, The crystal structure of indialite. CanadianMineralogist, Vol. 15, pp. 43–49.
Nassau K. (1984) Gemstone Enhancement. Butterworths,London.
Newly discovered beryl now classified as mineral (2003)Jewellery Net Asia, http://www.jewellerynetasia.com/news/view.jsp?tid=5795 (accessible with password), postedOctober 16.
Pezzotta F., Ed. (2001) Madagascar, a Mineral and GemstoneParadise. extraLapis English, No. 1, Lapis International LLC,East Hampton, CT, 100 pp.
Platonov A.N., Taran M.N., Klyakhin V.A. (1989) On two colourtypes of Mn3+-bearing beryls. Zeitschrift der DeutschenGemmologischen Gesellschaft, Vol. 38, pp. 147–154.
Rare pink-red beryl sparks market’s interest (2003) Jewellery NetAsia, http://www.jewellerynetasia.com/news/view.jsp?tid=5693(accessible with password), posted August 28.
Ristic S., Eichoff H.J. (1955) Les spectres d’absorption et réflexionde quelques variétés de béryls synthétiques et naturels.Congrés Groupement pour l’Avancement des Méthodsd’Analyse Spectrographiques des Produits Métallurgiques,Vol. 18, pp. 385–411.
Rossovskiy L.N. (1981) Rare-metal pegmatites with preciousstones and conditions of their formation (Hindu Kush).International Geology Review, Vol. 23, No. 11, pp. 1312–1320.
Schaller W.T., Stevens R.E., Jahns R.H. (1962) An unusual berylfrom Arizona. American Mineralogist, Vol. 47, pp. 672–699.
Schmetzer K., Berdensinski W., Bank H. (1974) Über dieMineralart Beryll, ihre Farben und Absorptionsspektren.Zeitschrift der Deutschen Gemmologischen Gesellschaft,Vol. 23, pp. 5–39.
Shatskiy V.S., Lebedev A.S., Pavlyuchenko V.S., Kovaleva L.T.,Koz’menko O.A., Yudin A.N., Belov N.V. (1981) Conditionsfor the entry of alkali cations into beryl. GeochemistryInternational, Vol. 18, No. 2, pp. 7–17.
Sherriff B.L., Grundy H.D., Hartman J.S., Hawthorne F.C., `CernyP. (1991) The incorporation of alkalis in beryl: MultinuclearMAS NMR and crystal-structure refinement study. CanadianMineralogist, Vol. 29, pp. 271–285.
Shigley J.E., Foord E.E. (1984) Gem-quality red beryl from theWah Wah Mountains, Utah. Gems & Gemology, Vol. 20, No.4, pp. 208–221.
Shigley J.E., McClure S.F., Cole J.E., Koivula J.I., Lu T., Elen S.,Demianets L.N. (2001) Hydrothermal synthetic red beryl fromthe Institute of Crystallography, Moscow. Gems & Gemology,Vol. 37, No. 1, pp. 42–55.
Shigley J.E., Thompson T.J., Keith J.D. (2003) Red beryl fromUtah: A review and update. Gems & Gemology, Vol. 39, No.4, pp. 302–313.
Sinkankas J. (1981) Emerald and Other Beryls. Chilton Book Co.,Radnor, PA.
Solntsev V.P., Bukin G.V. (1997) The color of natural beryls fromrare-metal Mozambique pegmatites. Russian Geology andGeophysics, Vol. 38, No. 10, pp. 1661–1668.
Staatz M.H., Griffitts W.R., Barnett P.R. (1965) Differences in theminor element compositions of beryl in various environ-ments. American Mineralogist, Vol. 50, pp. 1783–1795.
Warin R., Jacques B. (2003) Le béryl-Cs d’Ambatovita,Madagascar—Morphologie et aspects macroscopiques [TheCs-beryl of Ambatovita, Madagascar—Morphology andmacroscopic aspects]. Le Règne Minéral, No. 52, pp. 36–41.
Webster R. (1994) Gems—Their Sources, Descriptions andIdentification, 5th ed. Revised by P.G. Read, Butterworth-Heinemann, Oxford, England.
Wood D.L., Nassau K. (1968) The characterization of beryl andemerald by visible and infrared absorption spectroscopy.American Mineralogist, Vol. 53, pp. 777–800.
Zambonini F., Caglioti V. (1928) Ricerche chimiche sulla ros-terite di San Piero in Campo (Isola d’Elba) e sui berilli in gen-erale. Gazetta Chimica Italiana, Vol. 58, No. 131, pp.131–152.
PEZZOTTAITE FROM MADAGASCAR GEMS & GEMOLOGY WINTER 2003 301
